Machine Learning for COVID-19 Determination Using Surface-Enhanced Raman Spectroscopy

The rapid, low cost, and efficient detection of SARS-CoV-2 virus infection, especially in clinical samples, remains a major challenge. A promising solution to this problem is the combination of a spectroscopic technique: surface-enhanced Raman spectroscopy (SERS) with advanced chemometrics based on machine learning (ML) algorithms. In the present study, we conducted SERS investigations of saliva and nasopharyngeal swabs taken from a cohort of patients (saliva: 175; nasopharyngeal swabs: 114). Obtained SERS spectra were analyzed using a range of classifiers in which random forest (RF) achieved the best results, e.g., for saliva, the precision and recall equals 94.0% and 88.9%, respectively. The results demonstrate that even with a relatively small number of clinical samples, the combination of SERS and shallow machine learning can be used to identify SARS-CoV-2 virus in clinical practice.

SERS Signature of SARS-CoV-2 in Saliva and Nasopharyngeal Swabs: Towards Perspective COVID-19 Point-of-Care Diagnostics

In this study, the intrinsic surface-enhanced Raman spectroscopy (SERS)-based approach coupled with chemometric analysis was adopted to establish the biochemical fingerprint of SARS-CoV-2 infected human fluids: saliva and nasopharyngeal swabs. The numerical methods, partial least squares discriminant analysis (PLS-DA) and support vector machine classification (SVMC), facilitated the spectroscopic identification of the viral-specific molecules, molecular changes, and distinct physiological signatures of pathetically altered fluids. Next, we developed the reliable classification model for fast identification and differentiation of negative CoV(-) and positive CoV(+) groups. The PLS-DA calibration model was described by a great statistical value-RMSEC and RMSECV below 0.3 and R²_cal at the level of ~0.7 for both type of body fluids. The calculated diagnostic parameters for SVMC and PLS-DA at the stage of preparation of calibration model and classification of external samples simulating real diagnostic conditions evinced high accuracy, sensitivity, and specificity for saliva specimens. Here, we outlined the significant role of neopterin as the biomarker in the prediction of COVID-19 infection from nasopharyngeal swab. We also observed the increased content of nucleic acids of DNA/RNA and proteins such as ferritin as well as specific immunoglobulins. The developed SERS for SARS-CoV-2 approach allows: (i) fast, simple and non-invasive collection of analyzed specimens; (ii) fast response with the time of analysis below 15 min, and (iii) sensitive and reliable SERS-based screening of COVID-19 disease.

SERS-PLSR Analysis of Vaginal Microflora: Towards the Spectral Library of Microorganisms

The accurate identification of microorganisms belonging to vaginal microflora is crucial for establishing which microorganisms are responsible for microbial shifting from beneficial symbiotic to pathogenic bacteria and understanding pathogenesis leading to vaginosis and vaginal infections. In this study, we involved the surface-enhanced Raman spectroscopy (SERS) technique to compile the spectral signatures of the most significant microorganisms being part of the natural vaginal microbiota and some vaginal pathogens. Obtained data will supply our still developing spectral SERS database of microorganisms. The SERS results were assisted by Partial Least Squares Regression (PLSR), which visually discloses some dependencies between spectral images and hence their biochemical compositions of the outer structure. In our work, we focused on the most common and typical of the reproductive system microorganisms (Lactobacillus spp. and Bifidobacterium spp.) and vaginal pathogens: bacteria (e.g., Gardnerella vaginalis, Prevotella bivia, Atopobium vaginae), fungi (e.g., Candida albicans, Candida glabrata), and protozoa (Trichomonas vaginalis). The obtained results proved that each microorganism has its unique spectral fingerprint that differentiates it from the rest. Moreover, the discrimination was obtained at a high level of explained information by subsequent factors, e.g., in the inter-species distinction of Candida spp. the first three factors explain 98% of the variance in block Y with 95% of data within the X matrix, while in differentiation between Lactobacillus spp. and Bifidobacterium spp. (natural flora) and pathogen (e.g., Candida glabrata) the information is explained at the level of 45% of the Y matrix with 94% of original data. PLSR gave us insight into discriminating variables based on which the marker bands representing specific compounds in the outer structure of microorganisms were found: for Lactobacillus spp. 1400 cm⁻¹, for fungi 905 and 1209 cm⁻¹, and for protozoa 805, 890, 1062, 1185, 1300, 1555, and 1610 cm⁻¹. Then, they can be used as significant marker bands in the analysis of clinical subjects, e.g., vaginal swabs.

Dielectrophoresis-Based SERS Sensors for the Detection of Cancer Cells in Microfluidic Chips

The detection of freely circulating cancer cells (CTCs) is one of the greatest challenges of modern medical diagnostics. For several years, there has been increased attention on the use of surface-enhanced Raman spectroscopy (SERS) for the detection of CTCs. SERS is a non-destructive, accurate and precise technique, and the use of special SERS platforms even enables the amplification of weak signals from biological objects. In the current study, we demonstrate the unique arrangement of the SERS technique combined with the deposition of CTCs cells on the surface of the SERS platform via a dielectrophoretic effect. The appropriate frequencies of an alternating electric field and a selected shape of the electric field can result in the efficient deposition of CTCs on the SERS platform. The geometry of the microfluidic chip, the type of the cancer cells and the positive dielectrophoretic phenomenon resulted in the trapping of CTCs on the surface of the SERS platform. We presented results for two type of breast cancer cells, MCF-7 and MDA-MB-231, deposited from the 0.1 PBS solution. The limit of detection (LOD) is 20 cells/mL, which reflects the clinical potential and usefulness of the developed approach. We also provide a proof-of-concept for these CTCs deposited on the SERS platform from blood plasma.

Production of monodisperse drops from viscous fluids

Drops are often used as picoliter-sized reaction vessels, for example for high-throughput screening assays, or as templates to produce particles of controlled sizes and compositions. Many of these applications require close control over the size of drops, which can be achieved if they are produced with microfluidics. However, this tight size control comes at the expense of the throughput that is too low for many materials science and almost all industrial applications. To overcome this limitation, different parallelized microfluidic devices have been reported. These devices typically operate at high throughputs if the viscosity of the inner fluid is low. However, fluids that are processed into particles often contain high concentrations of reagents and therefore are rather viscous. We report a microfluidic device containing parallelized triangular nozzles with rectangular cross-sections that can process solutions with viscosities up to 155 mPa s into drops of well-defined sizes and narrow size distributions at significantly higher throughputs than what could be achieved previously. The increased throughput is enabled by the introduction of shunt channels: each nozzle is intersected by shunt channels that facilitate the backflow of the outer phase, thereby increasing the critical rate at which the fluid flow transitions from the dripping into the jetting regime. These modified nozzles open up new possibilities to employ drops made of viscous fluids as templates to produce particles with well-defined sizes for applications that require larger quantities.